Abstract
Guided by rules not always identifiable from the perspective of the present [2], most human cultures [3] have gone through periods of classificatory zeal and produced more than their fair share of ‘classes’ out of the objects, real and ideal, populating their universe. Western culture and its medicine have not been an exception and since Classical times, all manner of human ailments, including those pertaining to behaviour and/or the mind, have been assiduously classified [4].
One such classificatory drive, mainly concerning plants, animals and languages, appeared in the West during the 17th century [5]. By the end of the 18th century [6], it had been generalised to all Creation and crucially imbued with the belief that ‘classifying’ was inherent to man [7], who carried it out according to cognitive rules written either into his brain [8], mind [9] or language [10]. From then on, the debate was no longer about whether one should classify, but how [11]. It has remained so to this day. According to the majority, groupings have to be formed in terms of ‘privileged’ features [12] (e.g. the reproductive organs of plants). Linné and other 18th century botanists applied this method to diseases [13]. Towards the end of the 18th century, Adanson [14] dared to suggest that all features should be taken into account; his insight is now considered as a precursor of numerical taxonomy [15].
It is currently accepted that all classificatory acts must assume a ‘theory of concepts’ [16]. Eighteenth century classificators did likewise sharing the belief that concepts were ‘sortal’ instruments (i.e. devices to create classes by comparing candidate items with a paragon). The latter, in turn, was conceived as either a list of features (menu approach) or an image (a prototype) with sufficient discriminatory power to sort out all denizens of a finite universe into members and non-members of the given class [17]. The nature, origins and construction rules for the paragon were no longer discussed, for, by then, Locke's account of the meaning of terms (central to his theory of classification) had been widely accepted [18].
A second important 18th century assumption was that the ‘natural’ classificatory drive exhibited by man was innate (i.e. God-given). However, as evolutionary principles became established during the 19th century, the competing belief emerged that classifying was an ‘adaptative’ function (i.e. that it had been ‘selected in’ by nature for it conferred a cognitive advantage). All of these earlier assumptions are still embedded in current classificatory ideas, including those reigning in psychiatry.
In suchZeitgeist, classifying mental illness (or its phenomena) is mandatory, and as it was in 1798, in 1998 the mandate remains unchallenged. In rough and ready summary, views since the 19th century (which is the historical period to be dealt with in this paper) can be grouped according to whether they assume that the principles of ‘scientific’ c l a s sification are universal and: (i) apply to psychiatry
tout court [19]; (ii) apply to psychiatry mutatis mutandi [20]; or (iii) they do not apply at all, for so doing would be an instance of ignoratio elenchi[21, 22].
To the question, how it is to be decided which of these three views is the correct one, the conventional answer has been ‘by interrogating nature’ (i.e. by ‘empirical research’). However, this answer makes two unwarranted assumptions: that nature is a treasure of ready-made ‘knowledge’ [23], and that empirical research is ‘theory-free’. It is more likely, however, that at the moment the question has no answer for, so far, no science (including physics) has been able to develop a ‘crucial empirical test’ that can decide between rival conceptual systems. There is the additional complication that ‘empirical research’ is wedded to a view of nature according to which nature and her objects are fully independent from the mind, have a ‘mathematical structure’, etc., and hence it cannot be used to compare two different ‘epistemes’, particularly when one of them may actually reject that particular view of nature. In view of all this, to expect psychiatry to develop a crucial test is utopian (i.e. by empirical research alone it will not be able to identify, now or ever, a ‘real’ classification of mental disorders) [24].
A more realistic solution might, therefore, be for empirical work only to start once an agreement or convention has been reached between the various parties on matters such as: (i) the nature of the phenomena to be classified; (ii) the type of ‘concept’ and ‘class’ to be used; and (iii) a ‘sorting’ methodology. Because such agreement also requires that the conceptual analysis [25] be focused on a specific time and space frame (i.e. region of the world), it becomes obvious that a ‘conceptual history’ of psychiatric classifications is needed on which to base the agreement [26]. As far as I know, the latter has not yet been written. It is not that writings on ‘psychiatric classifications’ or their history do not abound, for since the 19th century the literature in this field has grown at much pace. It is that this material is surprisingly repetitious and parochial, and often intended to serve too many masters (e.g. clinical, administrative, didactic and propagandistic). Furthermore, they are based on inappropriate historiographical techniques and are not able to tell us about the history of concepts.
Nonetheless, these writings offer rich pickings. In addition to containing ‘facts’ about earlier classifications (in the manner of good catalogues), they tell something about the scientific and social world in which classifications were born. For example, during the 19th century, it was part of professional growth and success that alienists developed a personal classification [27]. These writings also tell about the deeper forces (social and otherwise) that have always controlled the methodology and implementation of classifications. Depending upon façon de parler, historians may say that such forces originate from a ‘paradigm’, ‘cognitive style’, ‘episteme’ or ‘fashion’. The advantage of the last term is that it, in opposition to the others, refers to a resistible social process, and hence removes the alibi that all alienists were controlled by ineluctable ‘cognitive rules’ [28].
By providing an outline for a future history of psychiatric classifications, this paper aims at helping readers escape from the conceptual myth that psychiatrists are obliged to classify mental disorders because that is the way in which the brain or language works. Because of its non-specificity (i.e. it can be marshalled in favour of any cause), this type of argument is unhelpful. Nonetheless, it is fair to say that classifications may serve a number of housekeeping and actuarial functions (e.g. providing a nomenclature, furnishing a basis for information retrieval, and description) [29]. In view of what will be discussed in this paper, it is less clear that they can also help with prediction and theory formulation. Little will be said here about the sortal act itself (for one, this raises the ugly problem of symptom and disease recognition) [30] or about the current products of classification.
Definitions now and in history
‘Classification’ is an ambiguous term in that it may refer to: (i) the creation of principles and rules in terms of which items, referents, features, and future groupings are to be constituted (what is called taxonomy when related to animals, plants and other objects of the tangible world) [31]; (ii) the act of grouping itself (sortal act; what in the case of psychiatry partially corresponds to diagnosis); and (iii) the products of activities (i) and/or (ii) (ICD-10 and other classificatory systems can be considered as an exemplar of this). There is no space in this paper even to start discussing the implications and functionalities attached to each of these meanings.
Let us start with a 20th century definition: ‘Classification is the identification of the category or group to which an individual or object belongs on the basis of its observed characteristics. When the characteristics are a number of numerical measurements, the assignment to groups is called by some statisticians discrimination, and the combination of measurements used is called a discriminant function’ [32].
There is little difference between the above and a 19th century definition: ‘In every act of classification, two steps must be taken; certain marks are to be selected, the possession of which is to be the title to admission into classes, and then all the objects that possess them are to be ascertained. When the marks selected are really important and connected closely with the nature and functions of the thing, the classification is said to be natural; where they are such as do not affect the nature of the objects materially, and belong in common to things the most different in their main properties, it is artificial … the advantages of classification are to give a convenient form to our acquirements, and to enlarge our knowledge of the relations in which different objects stand to one another…’ [33].
There now follows an 18th century view: ‘All [beings] are nuanced and change by degrees in nature. There is no one being that does not have either above or below it another that shares with him some features and that differs from him in other features’ [34]; and a 17th century one: ‘Nature, in the production of disease, is uniform and consistent; so much so, that for the same disease in different persons the symptoms are for the most part the same; and the selfsame phenomena that you would observe in the sickness of a Socrates you would observe in the sickness of a simpleton. Just so the universal characters of a plant are extended to every individual of the species; and whoever (I speak in the way of illustration) should accurately describe the colour, the taste, the smell, the figure, &c. of one single violet, would find that his description held good, there or thereabouts, for all the violets of that particular species upon the face of the earth’ [35]. The common points shown by these definitions reflect their common origin and hence the sharing of assumptions about both the entities to be classified and the rules for their classification.
Nomenclature
Any metalanguage (i.e. a second-order language) must include a minimum set of conceptual tools to allow the examination of the first-order language itself, in this case taxonomy (as opposed to the objects in the world to which the theory applies). A metalanguage for classificatory theory is gradually being constituted. The first distinction to be made concerns the classificatory system and its domain: the former refers to a frame, the set of rules and the definition of the basic units of analysis; the ‘domain’ concerns the population or universe of entities to be classified. These elements interact with one another but of particular relevance is the interaction between the definition of the units of analysis and its domain. This because many might say that the former helps to constitute the latter.
The metalanguage of classification has also been expressed in terms of a number of dichotomies: ‘categorical versus dimensional’; ‘monothetic versus polythetic’; ‘natural versus artificial’, ‘top-to-bottom versus bottom-up’, ‘structured versus listing’, ‘hierarchical versus non-hierarchical’, ‘exhaustive versus partial’, ‘idiographic versus nomothetic’, etc. These pairs of concepts criss-cross one another and it is not easy to deal with them independently. Some concern features or dimensions, others the compass of classifications, yet others relate to the methodology of taxonomy. Hence, only some may apply to psychiatry.
Ambiguity abounds within dichotomies. For example, in the dichotomy ‘categorical versus dimensional’, the first term can be interpreted as referring to the status of features or properties (a feature is either present or absent or 0 or 1) or to the style of sorting (i.e. an entity is or is not a member of a class); likewise the second term has a two-fold interpretation: that a feature is present as a continuum or that the entity (as defined as a cluster of dimensions) is only partially the member of a class. Categorical classifications (e.g. species and classes) are popular in biology [36], and seem to require that its exemplars be in steady state (at least within a given time frame).
Psychiatric phenomena, whether features (i.e. symptoms) or entities (i.e. diseases) do not meet this requirement and hence can be considered as ‘dimensional’. An important issue here concerns the problem of variations in either the quantity and/or the quality of the dimensions. If change occurred within a human time frame or range then it can be handled statistically. However, if it were to occur in an expanded time frame (say 100 000 years if due to a genetic mutation), then the change may be missed out altogether. This would be noticed after a recalibration is found not to tally the original classification. But then one would be confronted with the interesting situation that there are two ‘natural’ classifications of the same domain.
The ‘monothetic versus polythetic’ dichotomy concerns the number and predominance of the traits used to effect the sorting: monothetic definitions are based on the principle that classes consist of all the objects that take the same set of attributes. Polythetic definitions are based on the idea of ‘family resemblance’ and hence list many attributes or characters all of which are possessed by some members of the class, but none of which is possessed by all members of the class (i.e. each member just having some of the characters in question) [37].
Since the 18th century, classifications have been said to be ‘natural’ or ‘artificial’, but even this seemingly straightforward classification is ambiguous. According to a weak version, the terms characterise the old observation that objects can be classified in terms of either essential or man-made features (e.g. flowers can be grouped according to their sexual organs or simply in terms of their use, such as in funerals, weddings, etc.). According to the stronger version, however, the first horn of the dichotomy actually assumes the existence of ‘natural kinds’ [38]. ‘Natural classifications’ would thus be dictated by ‘reality’ itself whilst artificial ones would be manmade [39]. It goes without saying that those challenging the existence of natural kinds may want to argue that all classifications are, in fact, artificial.
‘Top-to-bottom’ and ‘bottom-up’ purportedly refer to the manner in which classifications are constructed. In practice, these terms overlap with the natural versus artificial dichotomy for they go back to whether classifications (as sets of categories) should be deduced from some high level theory or whether the furniture of the world should be painstakingly organised into groups which hopefully might become stable classes.
The ‘structured versus listing’ dichotomy concerns the internal organisation of the components of classifications. Good examples of pure ‘listings’ are the ICD-10 and DSM-IV glossaries which are best defined as partial inventories. Furthermore, the inclusion rules for these inventories are heterogeneous and originate from scientific and social sources. For example, some of the ‘clinical’ categories of DSM-IV are said to owe their inclusion more to pressures from the USA pharmaceutical or medical insurance industry than to ‘science’. Since social factors are always present in psychiatric classifications, this may or may not be a problem for DSM-IV and USA psychiatrists and it is nobody else's business. However, it becomes a problem when the glossary is exported lock-stock and barrel to other countries with an insurance industry structure totally different to that of the USA. ‘Structured’ classifications, in turn, are those whose categories together purport to offer a complete mapping of the universe to be classified (e.g. like the table of chemical elements). Although there were some during the 19th century, structured psychiatric classifications have not been seriously put forward during the 20th century.
The hierarchical versus one-level dichotomy also concerns the structure of classifications. Hierarchical classifications are decisional trees with higher classes embedding lower level ones. For example, Kraepelin's division of neurosis and psychosis, and the latter into schizophrenia and manic depressive insanity, is an example of a hierarchical dichotomous classification (said to have been suggested by Karl Kraepelin, Emil's older brother and distinguished classificator of molluscs) [40].
The ‘exhaustive versus partial’ dichotomy concerns the coverage or compass of a classification. Exhaustive classifications purport to cover all the entities in a given universe, hence, they tend to be deductive. For example, some 19th century classifications of mental disorders started by assuming that the entire realm of mental phenomena can be classified without residuum into intellectual, emotional and volitional. Then it was further assumed that any of these mental functions were independently affected by disease and it was thence inferred that there must be intellectual, emotional and volitional mental disorders. In fact, this is the basis for our current classifications leading to schizophrenia and paranoia (first group); mania, depression, anxiety, phobias, etc. (second group); and character and personality disorders (third group). Such classifications purported to cover all possibilities and anyone with a mental disorder had to fall into one of the groups. To deal with cases where more than one function was involved, rules of preference or hierarchies had to be created to decide on which ‘disease was primary’ [41].
Last, there is the occasionally mentioned idiographic versus nomothetic dichotomy whose historical origin has little to do with psychology or psychiatry although much to do with the classification of the sciences. This dichotomy is rich in meaning and hides a complex set of assumptions as to the nature of the natural and social sciences, as it was at the turn of the century. In 1894, in his inaugural lecture as Vice-Chancellor of Strassburg University, Wilhelm Windelband offered a summary of his views and focused on the difference between two types of sciences: ‘We can therefore say that in approaching reality the empirical sciences search for either of two things: the general in the form of natural laws or the special, as a specific event of history. They thus contemplate the permanent and immutable or the transitory as contained into real life happenings. The former sciences concern laws, the latter events; the former teach what has always been, the latter what has happened once. In the first case, scientific thinking is (if we were allowed to coin new technical terms) nomothetic, in the second case idiographic’ [42]. This distinction carries too much theoretical baggage to be applied simpliciter to the issue of group versus single case studies which are raging at the moment particularly in regards to classification and analysis in neuropsychology and neuroimaging studies. A study of Windelband's original reasoning, however, may throw some light on current confusions.
Concepts
It has been mentioned above that at the core of any taxonomy there is always a theory of concepts. This claim needs unpacking. Classifying can be made into a simple or a complex activity. According to the former, classifying is the act of sorting the entities of a given universe into pigeonholes provided by nature or invented by man. When analysed from a logical perspective pigeonholes have been called ‘classes’, and when studied as cognitive categories ‘concepts’. Now, while it is possible for anyone to create a classification without needing to explicitly sponsor a theory of ‘concepts’(ICD-10 and DSM-IV are good examples of this), it is also true that a theory of concepts is always implicitly assumed. In the latter case, the classificator will find that, because he has not worked out the theory of concepts he subscribes to, it might not be easy to understand some aspects of his own classification (e.g. its ‘truth-value’, internal structure, coverage, predictive capacity, etc.) [43].
However, good classifications must be more than concepts that are able to pigeonhole without error. Of all the potential benefits of a classification, the most important is its predictive power (i.e. its capacity to release additional knowledge about the entity at the very moment of its classification). The fact that this epistemological fruitfulness is rarely apparent in psychiatry needs to be accounted for. In this regard, one explanation may be that to be informative classificatory systems need to be structured, encompassing, and based on a general theory (like the table of chemical elements). For reasons which are discussed in other sections of this paper, there is little hope that this can be achieved in psychiatry.
The realisation that structured classifications and full categorical discrimination may be unachievable in psychiatry has led some to resort to other definitions of ‘class’ and ‘concept’ (e.g. ‘prototypes’) [44]. In this regard, Hampton has written: ‘it is claimed that uncertainty about classification is a result of people's inadequate knowledge of the categories that exist in the real world. The prototype view is directed, however, at a characterisation of exactly this inadequate knowledge — it is a model of the beliefs people have. Whether or not the real world is best described with a classical conceptual framework is an interesting and important question, but irrelevant to this psychological goal of the prototype model’ [45].
Is this relevant to psychiatry? Are the notions of ‘clearest case’, ‘best example’, and ‘procedural criteria’ applicable to this field? Can prototypes for all mental disorders be generated? What is their source? A plausible answer to the last question is that there are two sources: on the one hand, there is history (i.e. the received view of the disease), and on the other, there is ongoing clinical practice (i.e. the ‘lived experience’ of the maker of the prototype). How these two sources interact is one of the problems of psychiatry. For it is no good saying simpliciter that the latter provides the criteria to trim the former. It must not be forgotten that what history now gives us as ‘past’ was once upon the time the ‘lived experience’ of the creators of the first ‘prototype’. Now, as each successive generation trim and recalibrate the ‘received’ prototype new ones are being added. Faced with the historical series, we do not yet have rules to decide on which one should predominate.
Now, this problem has been dealt with by assuming that there is a linear progress in science and that per force the current prototype is best. Unfortunately, this assumption is untenable for in the specific case of mental disorders it is not right to say that all current trimmings and readjustments do is improve upon some ‘received’ prototype. The alternative view is to say that because they reflected exactly ‘lived clinical experience’, all prototypes are as valid as current ones. The problem here is caused by the fact that it is highly likely that through genetic mutation, medication, social pressures upon the language of description, fashions in the expression and formatting of subjective feelings, etc., the goal posts are moving all the time. In other words, the biological invariant (and its behavioural expression) are no more stable or reliable than a social invariant. Hence, it cannot be said that the latest prototype is necessarily the best in the temporal series.
The problems with prototypes, therefore, is that they do not really advance our knowledge about disease. They represent periodic recalibrations which remain valid within a given period of time, and there is no Platonic model to which they ineluctably take us. They simple capture, as Hampton put it, the extension of our ‘current beliefs’ about the object (disease) in question. In this regard, the hope that genetics will help psychiatry to fix the boundaries of mental disorders forever seems misplaced.
Assumptions
A number of assumptions are also made in regards to the feasibility of classifications. For example, as mentioned above, writers start from the assumption that classifying is the expression of an adaptative cognitive function. This assumption is expressed in various ways: for example, by using Kantian epistemology and affirming that man cannot stop classifying. In this sense, J.S. Mill also believed that classifying was built into grammar and that predicates are natural sortal mechanisms [46]. A ‘sortal concept’ is a ‘concept which conveys a criterion of identity and thus determines a type of object for which it makes sense to ask whether objects of the type are the same or different’ [47]. But not everyone in the 19th century agreed. For example, Durkheim believed that the human mind was not, in fact, very good at generating classifications and that it borrowed models from the symbolism available in society and culture. Hence, the so-called ‘logical’ categories were, in fact, social categories [48]. In this sense, classifying mental illness was more like classifying symbols than natural kinds.
But the same idea can also be approached from a psychological perspective, in relation to the role of both mental functions in general or to the role of human subjectivity. Thus, often a classificatory role has been attributed to ‘memory’ and considered: ‘essential to adaptive behaviour because it is organized in ways that make information gained from past experience applicable to present situations. And the essence of memory organization is classification. Although we experience only individual events, we remember them and identify occurrences as instances of classes or categories’ [49]. Others have proposed that there are at least two types of category learning: classification-based and inference-based [50]. So ingrained is this view, that it has been said that if language did not automatically perform a classificatory function, man's perception of the world would be compromised and he would live in a confused and confusing Jamesian mess.
Towards a history of psychiatric classifications
Foucault's suggestion that a new drive to catalogue nature started in the West during the 17th century is fundamentally correct. He further believed that this had to do with a new way of dealing with representation: namely, with the question of linking words to things (so that at a later stage one would only deal with words). For example, the crucial difference between 17th century animal histories and the medieval and Chinese bestiaries is that in the latter ‘everything was said about each animal’ whereas in histories all the old semantic content was cleaned out and systems of representation were created which applied to a variety of objects. The same changes affected history itself which from being a narrative became a technique to capture the truth of things. This forced history to catalogue and deal with the observed in a different way [51]. This is why during the 18th century, the issue of classifying became linked with the Condillacean concept of science as a ‘perfect language’ [52] and also with the debate on how sciences related to each other (the problem of the classification of the sciences) [53].
At the very end of the 18th century, the young Pinel [54] placed himself at the junction of two traditions. As one of the French translators of Cullen (who had admirably summarised the work of the continental nosographers) [55], Pinel declared himself a child of the 18th century; but in writing his Nosographie Philosophique [56] and challenging John Locke, he looked resolutely towards the new century. It was only by the division of the 19th century that Bouillaud, still influenced by Condillac, started to emphasise the value of creating a ‘perfect language’ for science, and the need for a stable nomenclature; in this latter regard he criticised Pinel for not having defined any of his key terms [57].
In his own definition of classification, however, Bouillaud remained an 18th century man: ‘a classification of diseases that is truly philosophical and rational must be based on knowledge on the nature of disease. All nosological buildings erected on any other foundations will remain fragile and lead to ruin’ [58]. (For a general criticism of this view see Riese [59].) In this short paper, there will only be space for setting the principles of a conceptual history. To this purpose, a specific historical event has been chosen for analysis, namely the 1860–1861 debate on classification at the Société Médico Psychologique. This case study will show that the said debate was a conceptual microcosm where all the main 19th century issues on psychiatric classification were ventilated.
Matters historiographical
There are at least three modes of writing on the history of psychiatric classifications: chronological catalogue [60], and social [61] and conceptual history. The last approach, to be followed in this paper, is based on the view that the history of words, concepts and the associated behaviours must be considered separately, and that for each mental symptom or disease a ‘convergence’ point can be found where the three elements come together. Conceptual history is based on the idea that mental symptoms are speech or communication acts conveying a biological signal which is heavily formatted by personal, social and cultural codes, and that upon being uttered these acts are additionally conceptualised by professional interpreters. Hence, the nature of psychiatric ‘reality’ is neither pure biology nor an empty social construct but a complex combination of the two. This general historiographical approach requires some modification to deal with psychiatric classifications. For example, the purpose, objective, conceptual assumptions, and social and temporal frames of classifications, what Lanteri-Laura has called ‘les références non-cliniques’ [62], need also to be identified.
The 1860–1861 SMPdebate
In 1934, Desruelles et al. offered the official view on the origin of French psychiatric classifications: ‘In 1843, the year when Annales Médico-Psychologiques started their publication, the classification of mental disorders generally accepted included mania, lypemania, monomania, dementia, paralytic insanity, and idiocy; to which some added stupidity. [In fact] this is the classification first put forward by Cullen in 1872 (mania, melancholia and dementia) to which Pinel added idiocy, Esquirol added monomania, and Georget added stupidity (i.e. states of confusional, demented and melancholic stupor which Esquirol had refused to consider as a separate entity). Lastly, Parchappe added paralytic insanity’ [63].
The view proposed by Desruelleset al. is factually [64] and historically telling. First, the saliency of Cullen's work has also been noticed by others (i.e. that by blending and streamlining 18th century disease classifications, the great Scotsman became a de facto culture carrier) [65]. Second, because 20th century medicine can be considered as sharing the same classification principles as the previous two centuries, it could be said that Cullen's taxonomic ideas have more importance than what has hitherto been considered. Third, the fact that all leading European psychiatric nosologies (e.g. German, French and Italian) seem to have the same origin (i.e. Cullen's nosology) suggests that later divergences may have resulted more from the growth of European nationalisms than from any logical or empirical developments within psychiatry itself [66].
While during the 19th century most alienists happily accepted the view that classifying was an essential aspect of their work, writers with a deeper social understanding were more ambivalent. A good example is Philippe Buchez [67], who satirised: ‘Upon believing that they have completed their studies, rhetoricians will compose a tragedy and alienists a classification’ [68]. However, he also wrote: ‘in addition to facilitating teaching and helping to remember — by themselves important functions [classifications] have as their most important objective to carry out a diagnosis, now called differential diagnosis. Now since diagnosis is at the basis of treatment, then it can be said that in the last analysis the objective of classifications is treatment’ [69]. This feeling of confusion about psychiatric classifications continued till the end of the century. Thus, Féré wrote: ‘those who debate on psychiatric classifications sound like the workers at the Tower of Babel: the more they talk the less one understands them. If the terms do not mean the same for everyone then they run the risk of being applied to different clinical states’ [70]. This trouble with definitions led B a i l l a rger to suggest that the Société Médico-Psychologique should ‘fix the meaning of the main forms of mental disorder so that their scientific path could, so to speak, be controlled’ [71].
Desruelles et al. also reported that 19th century French alienists used at least eight criteria to classify mental disorders: the cause of the disorder (aetiological), the substratum of the disorder (e.g. anatomy), the clinical outcome (whether curable or incurable), actuarial (according to what was observed in the statistics of the main asylums in France), phenomenological (according to whether or not the disorder included delusions), ‘natural’ (i.e. whether it corresponded to ‘real types’ as given in nature), psychological (i.e. what mental faculty was assumed to be impaired), and disease course.
With all these approaches vying for power, it is not surprising that the Société Médico Psychologique resorted to a debate [72] on psychiatric classifications.
The debate
The debate was started on 12 November 1860 by Delasiauve [73] who, on the excuse of dealing with Buchez's positive review of a book recently published by Morel [74], launched forth into an analysis of the classificatory ideas of Esquirol, Ferrus, Falret, Girard de Cailleux, Lasègue and Baillarger. Of Morel's aetiological classification, Delasiauve wrote: ‘Morel is evidently intoxicated by his views. What seems to have seduced him and to a certain extent Buchez is the disorders caused by alcohol, lead and epilepsy where the cause is tangible and its effects understandable. Not so with other states where there are multiple factors and influences’ [75]. Buchez replied briefly: ‘Mental illness is characterized by signs and symptoms and these have served always as classificatory principles. Are they sufficient? The answer is not … it is necessary to search for the pathogenesis of mental disorder … to describe mental illness is not to classify it’ [76].
The debate on the Session of 26 November 1860 was started by Jules Falret [77], who offered to identify ‘the principles that govern, in all sciences, the development of natural classifications’: ‘1) A class should be defined in terms of a set of features present in all the objects to be encompassed and not in terms of one character that might artificially bring together objects which would be different if other features were to be considered; 2) the said feature-set should itself be organized in a hierarchy so that its essential components are clearly identified; 3) the objects that come under one class should not only share the feature set at a given moment in time but show that they have evolved in a similar manner [acquired the features in an order that can be predicted]’ [78]. It can be clearly seen that criterion 3) reflects the influence of evolution theory and sets a task that remains unfulfilled to this day: few psychiatrists will consider the order in which the symptoms appear as a classificatory criterion [79].
Falret then proceeded to criticise all available classifications for they rely on one feature or character alone, for example the involvement of: ‘1) an intellectual faculty or 2) a predominant idea or emotion or 3) act or 4) the features and extension of a delusion’ [80]. Based on this, he went on to conclude that ‘mania, monomania, melancholia and dementia are but provisional symptomatic clusters and not true natural species of mental disorders’ [81].
Morel had attended the session of 26 November and ‘although [he] was not planning to talk … he felt the need to intervene as his ideas had been questioned by Delasiauve and Falret’. Not having been prepared, his speech was rambly and concentrated on justifying his view that hereditary mental disorders existed although ‘three elements converge to create them: predisposition, a cause, and a series of transformations of pathological phenomena that … determine the place that a given disorder will occupy in the nosological classification’ [82].
The session of 10 December was taken by the debate between Adolphe Garnier [83] and Alfred Maury [84]. Garnier started by suggesting that the dichotomy ‘natural–artificial’ should be replaced by ‘essential–superficial’. Since it was impossible to take all the features of the objects to be classified into account, groupings based on a limited number of characters always led to blurred boundaries and hence ‘there were always patients who floated between two classes’. Geoffrey Saint-Hilaire and Cuvier had had a clash on this very issue a long time ago and ‘there was no solution to it’. This was the reason why alienists tried to complement their classifications with speculation about aetiology but the latter should not be considered as a feature of the objects in question. Garnier finished by suggesting an ‘essential classification of mental disorders based on mental faculties’ [85].
Maury rose to say that since science had not advanced sufficiently, only artificial classifications were possible, particularly in the area of psychological medicine where little was known about causes. He attacked Garnier for sponsoring a ‘psychological view’ of mental illness, for considering that ‘it is a consequence of an emotional turmoil already present in the heart of man, and which in the event enslaves him and takes his freedom away’. While strong emotions might on occasions cause insanity, there were many cases where the symptoms overcame the patient (e.g. those resulting from brain diseases forced changes in his behaviour). The question ‘was not to classify mental disorders from a philosophical viewpoint in terms of which mental faculty seemed involved but in terms of its pathological origins … otherwise it will not be possible to differentiate mental illness from normal behaviour (e.g. monomania from dreaming)’. ‘We must not forget that we are not here in the world of metaphysics but in psychological medicine. We classify mental disorders to cure them and this is why we must try and find out their aetiology’ [86].
The next session, on 24 December, was taken by Garnier's rejoinder. After iterating his view that ‘looking for efficient causes according to Bacon's principles’ [87] was the best way to achieve a classification, he explained why Delasiauve had felt unable to classify insanity according to the received view on the psychology of the intellectual functions: ‘it is not possible to create a classification of insanity based on a conventional division of intelligence into judgement, reasoning, etc. because the latter modes are not independent but blend with each other and with attention and memory … Hence, it might be better to divide intelligence according to the objects it deals with’ [88].
Garnier then turned on Maury: ‘He tells me, you do not take into account physical causes but are only preoccupied with psychological ones … I reply, although a spiritualist I am not afraid of matter but when I look for it I cannot find it … insanity is a disturbance of the intellectual faculties and requires a specific method of observation … there is as little point in listing insanities due to saturnism, alcoholism, etc. as there is listing others due to diseases of the liver. Changes in this organ do not cause the insanity. It is in the brain where the cause must be sought. If that part of the brain that produces hallucinations is lesioned in any way, including by the touching with a finger, it will lead to the same symptom’ [89].
The following session took place on 29 January 1861, and its protagonists were Buchez and Garnier. The former started by producing a summary of what had gone on so far: ‘on the one hand, papers have been read here suggesting that insanity would be just the one disease with diverse manifestations’ [90] from which follows that descriptive and classificatory publications only scratch its surface and ‘on the other hand, a paper has been read trying to identify organic aetiology, pathogenesis’… ‘the Baconian method is used in medicine under the name of “method of exclusion” but has had little influence on natural history’; ‘Diseases are not independent entities, existing by themselves, with a life of their own, like plants or animals. They fully depend upon a living organism and exist only in it’ [91]. Because ‘a classification must be above all a faithful reflection of the science of its time’, Buchez felt that he needed to explain to the audience what was the ongoing state of the philosophy of science and proceeded to do so. He concluded by, once again, giving support to Morel and his organic classification [92].
Garnier was surprisingly conciliatory in his intervention: ‘The question then is to know whether there are various insanities or only one. Buchez seems inclined towards the latter. To determine this we must use the Baconian method … But he is right in saying that the latter is not relevant to natural history’ [93]. The session was closed by the blunt (and correct) observation of Dr. Archambault's [94]: ‘Garnier and Buchez seem to be employing the words madness (folie), mental alienation, mental illness etc. as if they meant the same. As far as I am concerned they have different meaning’. Stung by the comment Buchez rose to his feet: ‘madness and mental alienation do not mean the same; mental alienation has however a much wider meaning’ [95].
The debate went on for three more sessions: 25 February, 25 March and 29 April. By then it was repetitious and although new speakers appeared (Parchappe and Lisle) no new conceptual point was made.
In summary, the 1860–1861 French debate on psychiatric classifications is important for it was a microcosm where all the important issues of the conceptual and empirical difficulties involved in this field were made. Desruelles et al. saw it as reflecting a clash between the traditionalists (defenders of Esquirol's monomania concept) and innovators (those who followed Morel) [96].
Although historically correct, on hindsight this interpretation is found too narrow, and its wider implications need teasing out. Salient is the fact that the debate, although taking place in a Medico-Psychological Society founded by alienists, and apparently starting from a practical problem, was soon taken over by philosophers of mind. This meant that there was the need to sort out principles and rules, and to reach the agreement that empirical research alone could not solve the problem. Notable also is the fact that all participants at the time soon realised that classifying diseases was not like classifying plants or animals and that the conceptual problem involved was completely different. Thus, and in terms of the classification of points of view suggested at the beginning of this paper, French alienists believed that general taxonomic principles only applied mutatis mutandi. They also realised that the question of whether there was only one or many forms of mental disorder (i.e. the ‘unitary psychosis’ issue) was of central import to the classification debate and that this question was not necessarily empirical in nature. Last, no clear decision was taken as to whether psychological (descriptive) or organic (aetiological) criteria should be used. In general, although lip service was paid to organic classifications, it was agreed that so little was known about aetiology that such approach remained utopian.
The 19th century French alienists were well aware that confusion about classification was an obstacle to research, and in 1889 a second debate took place at the Société Médico-Psychologique [97]. At the European level, there were also two meetings. In 1885, the Congress of Mental Medicine at Antwerp appointed a Commission to consider all existing classifications. This was discussed at the Paris Congress of 1889 and a classification drawn by a Dr. Morel from Ghent was adopted. Rather ruefully, Daniel Hack Tuke wrote in 1892: ‘it has yet to be seen whether asylum physicians will adopt it in their tables’ [98].
Conclusions
For the conceptual historian, there are basically two ways of approaching psychiatric classifications. One is to write from within the episteme and accept both the ‘received view’ that classifying is inherent to the human mind and the inference (yet to be tested) that this means that psychiatrists are compelled to exercise this function in their field of interest. According to this first approach, all the historian can do is catalogue (once again) the products of earlier classificatory enterprises. As shown above, this type of work has been done ad nauseam.
The second approach is to study psychiatric classifications from outside the episteme, not to take any views for granted, and consider psychiatric classifications as cultural products. This should also include examining the scientific evidence that brains, minds or languages actually possess built-in classificatory algorithms. In practice, however, it does not matter whether or not they do for even if their existence is granted, it does not follow that their possession should drive men to classify any and all objects they come across. In fact, it can be assumed that the classificatory capability was ‘selected in’ by evolution to allow for the rapid classification of plants, animals, stones, and other objects relevant to man's survival. It remains to be proven: (i) whether this capability extends to all objects in man's new environments; and (ii) if it does, whether it is in his best interest (and creating science is considered as one of them) to classify ideal objects, hypothetical constructs and other entities whose real management and understanding may require the use of a personalised semantics.
This paper has been written from the perspective of the second approach. This is the reason why its first long section was dedicated to mapping the conceptual structure of the various components of the classificatory activity and to linking them, as far as it was possible in terms of the available scholarship, to the context in which they were first thought out. The first conclusion here is that there are major gaps in the literature and that much work needs to be done before a coherent picture emerges. The second conclusion is that conceptual history and straight analytical philosophy work well together in this field. A third conclusion is that even after showing that there are weighty conceptual reasons why psychiatric classifications are not working, there is no reason to infer from that that classifying in psychiatry is a useless exercise. All that follows is that much more conceptual work needs to be done if we are going to develop classifications which have more than an actuarial function.
Since medicine is not a contemplative but a modificatory activity, a fourth conclusion is that what is required is classifications that can release new information about the object just classified. The received model of mental (a variant of the old anatomo-clinical hypothesis) still enjoins us to anchor all ‘essential’ classificatory criteria in neurobiology. Looking for a ‘biological invariant’ responsible for surface events (symptoms) seems a task worth pursuing. But it also looks as if, in order to really understand the role of the biological invariant, some naive beliefs as to the stability, duration and reliability of biological substrata will have to be given up. History seems to suggest that the biological substratum is ‘invariant’ only within a given time frame. We know little about the latter, particularly in regards to the genes related to the main mental disorders.
Nonetheless, the view that mental disorders are behavioural epiphenomena wrapped around molecular changes is rapidly developing in some quarters. According to this view, a ‘natural’ classification should develop once all responsible genes have been identified. What is more, it is also expected that such classification might even help with the characterisation of stable phenotypes. The question here is whether such a molecularly based classification would actually be a classification of mental disorders. Many may feel that it would not for molecular changes are not yet a mental disorder, and because the latter is intrinsically a complex construct that must include the pertinent formatting codes.
But the fact that genetics alone will not do, should not stampede psychiatrists into searching for a social invariant. The belief that all mental disorders are merely social constructs is not threatening to psychiatry because it undermines the professional existence of psychiatrists; it is threatening for it does not offer the steadiness required to create a predictive system, which is what medicine is all about.
Taxonomy and its associated classificatory activities constitute a self-contained and more or less exhaustive conceptual system. This means that within a given historical period, thinking about and crafting classifications is like playing a game of chess in that everything will occur within strict boundaries and according to explicit or implicit rules. For example, not all possible moves will be made: some because they are forbidden by the rules, others because they are patently suicidal, and yet others because they are not fashionable. The same with classifications. For example, the concept of class we have inherited from 18th century botany still encourages psychiatrists in the 20th century to dream of classifyingper genus et differentia, even if this does not apply to mental disorders at all. It was precisely to illustrate this point, namely, the constrained conceptual environment in which all talk on classification takes place, that a detailed analysis of the 1860–1861 French debate is offered in this paper. Noticeable was the fact that whether psychologist, historian, philosopher or alienist, the participants in the debate agreed on fundamentals such as that there is a need to classify, that mental disorders are susceptible to classification, that classes must have clear cut boundaries, etc.
The same can be said of the current literature on psychiatric classifications. Although increasingly voluminous and on occasions imaginative, it is still postulated by ‘unsaids’ and with very rare exceptions does not allow itself to think the unthinkable; that is, psychiatric classifications may be called into question because not enough is known about aetiology, or because mental illness is so complex and mysterious that it cannot be classified, or because not enough empirical research is being done, or because one is not being scientific enough and allows social factors to contaminate what should be a purely surgical enterprise (carving nature at the joints). But rarely if ever the literature challenges the validity of the classificatory act itself, irrespective of whether mental disorders are or are not susceptible to classification. It is one of the contentions of this paper that that conceptual (chess) move needs to be explored.